Inspection system of listeria monocytogenes biofilm formation on conveyer belt

Abstract

This thesis proposes a novel method to enhance out-of-focus regions in microscopic images and detect Listeria Monocytogenes(L. monocytogenes) in biofilm on a conveyor belt from an image with the aim to count the colonies of L. monocytogenes. Consider that a microscope can focus on a single point in the sample, this limitation makes each image contain both in-focus and out-of-focus regions. These out-of-focus regions are blurred with different blurring scale, thus the proposed thesis technique uses nonlinear piecewise least-square curve-fitting filter and the gradient to enhance the out-of-focus regions. Finally, watershed algorithm and morphological operations are used to segment and count the colonies. The proposed thesis was evaluated on a set of 30 L. monocytogenes microscopic images and a set of 90 synthetic test images. The counting results reveal that our proposed method can improve the quality of an image by enhancing the out-of-focus regions while preserving other regions which leads to more accurate counting results.